← 研踪 PaperTrailarXiv:2610.02545 · v2
Reward Inflation: A Healthy Stimulus for Reinforcement Learning
首次提交 2026-10-02 06:33
版本更新 2026-11-12 21:51
雷达观测 2026-11-13 14:15
摘要
Reward serves as the primary learning signal in reinforcement learning (RL). However, while reward magnitudes are typically held fixed throughout training, their temporal modulation remains underexplored. In this paper, we propose reward inflation, a gradual scaling of rewards over the course of training, and show that it can act as a healthy stimulus for RL. Theoretically, reward inflation induces an implicit recency weighting that upweights recent transitions during policy updates, enabling faster adaptation. We further show that reward inflation suppresses the emergence of dormant neurons by sustaining gradient signals as the policy saturates, suggesting a possible contribution to plasticity preservation. Empirical results on ALE games and MuJoCo tasks corroborate these findings, showing that an appropriate level of reward inflation benefits a broad range of tasks. Finally, we introduce Fed, an adaptive variant that adjusts the inflation level on the fly, and find that it often improves upon fixed inflation. Source code is available at https://github.com/northadventure/RewardInflation.
作者备注：Accepted at NeurIPS 2026
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v2 · 2026-11-12 21:51 · 最新观测版本
Reward Inflation: A Healthy Stimulus for Reinforcement Learning
Reward serves as the primary learning signal in reinforcement learning (RL). However, while reward magnitudes are typically held fixed throughout training, their temporal modulation remains underexplored. In this paper, we propose reward inflation, a gradual scaling of rewards over the course of training, and show that it can act as a healthy stimulus for RL. Theoretically, reward inflation induces an implicit recency weighting that upweights recent transitions during policy updates, enabling faster adaptation. We further show that reward inflation suppresses the emergence of dormant neurons by sustaining gradient signals as the policy saturates, suggesting a possible contribution to plasticity preservation. Empirical results on ALE games and MuJoCo tasks corroborate these findings, showing that an appropriate level of reward inflation benefits a broad range of tasks. Finally, we introduce Fed, an adaptive variant that adjusts the inflation level on the fly, and find that it often improves upon fixed inflation. Source code is available at https://github.com/northadventure/RewardInflation.
查看此版本解读 →查看此版本原文 ↗